nvidia/nemotron-3-super-120b-a12b behavioral LLM fingerprint
This page breaks down the exact reference that VerifyLLMAPI uses for nvidia/nemotron-3-super-120b-a12b: 111 normalized answers across eight low-entropy probes, plus its nearest enrolled models by mean Jensen–Shannon divergence.
Fingerprint summary
- Reference answers
- 111
- Samples per cell
- 11–15
- Nearest model
- qwen/qwen3.6-35b-a3b
- Nearest mean JSD
- 0.617479
The most concentrated cell was animal-random|zh: 企鹅 held 46.7% of its answers and entropy measured 2.197 bits. The broadest cell was animal-random|ar at 3.507 bits.
Against the closest enrolled model, animal-random|en contributed the largest cell distance at JSD 1.000000. This is a descriptive comparison, not proof that either API route serves a named checkpoint.
Completeness: the author dataset allows up to 15 valid answers per selected cell in reference half A. This model contains 111/120 across the eight cells. We show the actual denominator for every share.
The eight test requests
The research pipeline sent each user prompt with a one-answer system constraint, temperature 1, a 16-token output cap, and reasoning disabled. VerifyLLMAPI uses the same prompt text and normalization target.
| Cell | Task | Exact request text | Normalization and n |
|---|---|---|---|
animal-random|en |
Random animal · English | Name a random animal.System constraintAnswer with a single word or a single number only. Output nothing else: no explanation, no punctuation, no quotation marks. |
word · 14/15 reference samples |
word-random|en |
Random word · English | Name a random word.System constraintAnswer with a single word or a single number only. Output nothing else: no explanation, no punctuation, no quotation marks. |
word · 11/15 reference samples |
animal-random|ru |
Random animal · Russian | Назови случайное животное.System constraintОтвечай только одним словом или одним числом. Не выводи ничего больше: без объяснений, без знаков препинания, без кавычек. |
word · 15/15 reference samples |
animal-random|ar |
Random animal · Arabic | اذكر حيواناً عشوائياً.System constraintأجب بكلمة واحدة أو رقم واحد فقط. لا تكتب أي شيء آخر: لا شرح ولا علامات ترقيم ولا علامات اقتباس. |
word · 15/15 reference samples |
letter-random|zh |
Random character · Chinese | 说出一个随机的汉字。System constraint只用一个词或一个数字回答。不要输出任何其他内容：不要解释，不要标点符号，不要引号。 |
grapheme · 14/15 reference samples |
animal-random|zh |
Random animal · Chinese | 说出一个随机的动物。System constraint只用一个词或一个数字回答。不要输出任何其他内容：不要解释，不要标点符号，不要引号。 |
word · 15/15 reference samples |
letter-random|ar |
Random letter · Arabic | اذكر حرفاً عشوائياً من الحروف الأبجدية.System constraintأجب بكلمة واحدة أو رقم واحد فقط. لا تكتب أي شيء آخر: لا شرح ولا علامات ترقيم ولا علامات اقتباس. |
grapheme · 12/15 reference samples |
color-random|zh |
Random color · Chinese | 说出一个随机的颜色。System constraint只用一个词或一个数字回答。不要输出任何其他内容：不要解释，不要标点符号，不要引号。 |
word · 15/15 reference samples |
Response distributions
These are normalized reference-half counts from the published OpenRouter dataset. The tables contain aggregate values, not user data or a live VerifyLLMAPI scan.
1. animal-random|en
Random animal, English. 14 valid reference answers produced 8 normalized values. The mode was axolotl at 7/14 (50.0%); entropy was 2.404 bits.
| Normalized response | Count | Share |
|---|---|---|
axolotl |
7 | 50.0% |
crab |
1 | 7.1% |
flamingo |
1 | 7.1% |
fox |
1 | 7.1% |
kangaroo |
1 | 7.1% |
octopus |
1 | 7.1% |
owl |
1 | 7.1% |
tarsier |
1 | 7.1% |
2. word-random|en
Random word, English. 11 valid reference answers produced 9 normalized values. The mode was apple at 2/11 (18.2%); entropy was 3.096 bits.
| Normalized response | Count | Share |
|---|---|---|
apple |
2 | 18.2% |
banana |
2 | 18.2% |
bloop |
1 | 9.1% |
blue |
1 | 9.1% |
dragonfly |
1 | 9.1% |
luminescent |
1 | 9.1% |
luna |
1 | 9.1% |
sunset |
1 | 9.1% |
zinnia |
1 | 9.1% |
3. animal-random|ru
Random animal, Russian. 15 valid reference answers produced 12 normalized values. The mode was коала at 3/15 (20.0%); entropy was 3.457 bits.
| Normalized response | Count | Share |
|---|---|---|
коала |
3 | 20.0% |
крокодил |
2 | 13.3% |
гепард |
1 | 6.7% |
жираф |
1 | 6.7% |
зебра |
1 | 6.7% |
капибара |
1 | 6.7% |
кенгуру |
1 | 6.7% |
киви |
1 | 6.7% |
лев |
1 | 6.7% |
лис |
1 | 6.7% |
лисёнок |
1 | 6.7% |
слон |
1 | 6.7% |
4. animal-random|ar
Random animal, Arabic. 15 valid reference answers produced 12 normalized values. The mode was النمر at 2/15 (13.3%); entropy was 3.507 bits.
| Normalized response | Count | Share |
|---|---|---|
النمر |
2 | 13.3% |
زرافة |
2 | 13.3% |
نمر |
2 | 13.3% |
أرنب |
1 | 6.7% |
البطريق |
1 | 6.7% |
الكوالا |
1 | 6.7% |
بومة |
1 | 6.7% |
ثعلب |
1 | 6.7% |
ضفدع |
1 | 6.7% |
فيل |
1 | 6.7% |
قرد |
1 | 6.7% |
قطة |
1 | 6.7% |
5. letter-random|zh
Random character, Chinese. 14 valid reference answers produced 12 normalized values. The mode was 龙 at 3/14 (21.4%); entropy was 3.468 bits.
| Normalized response | Count | Share |
|---|---|---|
龙 |
3 | 21.4% |
山 |
1 | 7.1% |
春 |
1 | 7.1% |
曜 |
1 | 7.1% |
木 |
1 | 7.1% |
桜 |
1 | 7.1% |
水 |
1 | 7.1% |
熔 |
1 | 7.1% |
电 |
1 | 7.1% |
福 |
1 | 7.1% |
艳 |
1 | 7.1% |
艹 |
1 | 7.1% |
6. animal-random|zh
Random animal, Chinese. 15 valid reference answers produced 6 normalized values. The mode was 企鹅 at 7/15 (46.7%); entropy was 2.197 bits.
| Normalized response | Count | Share |
|---|---|---|
企鹅 |
7 | 46.7% |
斑马 |
2 | 13.3% |
猫 |
2 | 13.3% |
长颈鹿 |
2 | 13.3% |
大象 |
1 | 6.7% |
狐狸 |
1 | 6.7% |
7. letter-random|ar
Random letter, Arabic. 12 valid reference answers produced 6 normalized values. The mode was م at 5/12 (41.7%); entropy was 2.221 bits.
| Normalized response | Count | Share |
|---|---|---|
م |
5 | 41.7% |
m |
3 | 25.0% |
j |
1 | 8.3% |
k |
1 | 8.3% |
ب |
1 | 8.3% |
ج |
1 | 8.3% |
8. color-random|zh
Random color, Chinese. 15 valid reference answers produced 7 normalized values. The mode was 蓝色 at 4/15 (26.7%); entropy was 2.606 bits.
| Normalized response | Count | Share |
|---|---|---|
蓝色 |
4 | 26.7% |
绿色 |
3 | 20.0% |
蓝 |
3 | 20.0% |
紫色 |
2 | 13.3% |
blue |
1 | 6.7% |
红 |
1 | 6.7% |
青色 |
1 | 6.7% |
Three nearest enrolled references
We compute base-2 JSD for each of the eight categorical distributions, then average the eight distances. A lower value means more similar answer frequencies within this product reference. It does not identify an unknown route by itself.
| Rank | Reference model | Mean 8-cell JSD |
|---|---|---|
| 1 | qwen/qwen3.6-35b-a3b |
0.617479 |
| 2 | mistralai/mistral-medium-3-5 |
0.638411 |
| 3 | inclusionai/ling-2.6-flash |
0.651058 |
How many live requests does verification use?
| Profile | Cells × samples | Fresh model requests | Same-data EER |
|---|---|---|---|
| Quick, default complete check | 4 × 5 | 20 | 11.8% |
| Standard | 8 × 5 | 40 | 8.1% |
| Paper budget | 8 × 15 | 120 | 6.7% |
The current-Agent workflow starts with quick. Each answer comes from a fresh Codex or Claude Code process that reuses host-managed login. These requests can consume host allowance or provider balance.
What this fingerprint can and cannot show
- The reference records behavior observed through OpenRouter in the author dataset. Provider wrappers, model updates, decoding, and date can move a distribution.
- The author collection set temperature 1 and disabled reasoning. Current-Agent mode labels sampling as a host default because supported Agent CLIs expose no temperature flag.
- The eight high-gap cells and product thresholds use the same public cohort. Their EER values are calibration results, not an independent production accuracy claim.
- A nearest reference is a similarity result. VerifyLLMAPI returns INCONCLUSIVE when the claimed route lacks an enrolled fingerprint.
Primary sources, license, and reproducibility
- Tomas Bruckner, One Token Is Enough: Fingerprinting and Verifying Large Language Models from Single-Token Output Distributions, arXiv:2607.10252v1.
- Author dataset and derived results, DOI 10.5281/zenodo.21278557, CC BY 4.0.
- Author collection and analysis software, DOI 10.5281/zenodo.21278793, MIT.
- VerifyLLMAPI atlas JSON, built from reference
pamela-openrouter-2026-07-8cell-ref-a, SHA-25651f3f63cbcfb56b151055d2d10e36a9cc2f0644cc68b580784853d68951fa866.
Page evidence SHA-256: 432980e6687ca7e75e89aa6dfa6aedd387c8439d160e8298294076087b38c9ee. This hash covers the model ID, eight cell metrics, nearest references, and strongest cell difference in the generated atlas.